Home / How redaction actually works
How redaction actually works
Nearly every free "redact" tool draws a black rectangle over the text and calls it done, leaving the original text sitting in the file underneath, ready for anyone to copy and paste. This page explains, plainly, why that is not redaction, and exactly what this editor does differently.
Find & Cover, the redaction tool in this editor, actually deletes the matched text from the PDF’s own content — not a visual layer drawn on top of it — and every export independently re-verifies the removal and shows you the result. The one stated exception: a rotated page is covered but not yet removed, and the tool says so on screen at the moment it applies.
The trap almost every "redaction" tool sets
A PDF page is not a flat picture. It is a set of instructions — draw this text here, draw this shape there — layered on top of each other in the order they were added. When a tool "redacts" something by drawing a black rectangle after the text, all it has done is add one more instruction: draw a black shape over that spot. The text instruction underneath is completely unchanged. Anyone can select it, copy it, search for it, or simply open the file in a viewer that lets them peel back or ignore that top layer.
This is not a rare or theoretical failure. Government agencies, courts, and law firms have published documents redacted exactly this way, and had the "removed" names, figures, or classified passages recovered in minutes by a member of the public who knew to try selecting the text underneath the black box.
What actually happens when Find & Cover redacts something here
Find & Cover searches a page for a pattern — an email address, a phone number, a Social Security number, an ID number, or anything you type in yourself — and lets you confirm which matches to act on. Once confirmed, exporting the document does two distinct things, deliberately, not one:
- It draws an opaque cover over the matched area, so the exported page still reads as visibly redacted rather than blank or broken — the same reason official, properly redacted documents show a black bar rather than empty white space.
- Separately, and this is the part almost nobody else does, it opens the page’s own underlying content — the actual instructions that make up the page — finds the specific text-drawing instructions inside the redacted area, and deletes them outright. Not blanked, not hidden, not marked invisible: removed from the instruction list entirely, the same as if that word had never been typed onto the page.
What is left in the file, after export, is a page whose instructions no longer contain the text you redacted — alongside every other word on that page, untouched, exactly as it was.
How to verify this yourself, not just believe it
A claim like this should not need to be taken on trust, so it is not left as one:
- The export report proves it independently. After exporting a redacted document, the verification screen re-opens the file with a second, separate engine and re-extracts its text — the same check already applied to page count and file structure for every export — and states plainly whether the redacted content is actually gone.
- Try it in a completely different program. Export a redacted file, open it in any other PDF reader, and try to select or search for the text you redacted. There is nothing there to find, because there is nothing there.
The one honest exception
On a page that has been rotated, the redaction still covers the matched area visually, but the underlying text is currently left in place rather than removed. Doing this correctly on a rotated page means converting the redaction box back into the page’s own, un-rotated coordinate space — real, additional work that has not been done yet, and shipping a version that might convert that incorrectly and remove the wrong area would be worse than shipping nothing for that one case. The redaction tool says this on screen, at the moment it would actually matter, not in a policy page you would have to already know to go looking for. If a rotated page needs a genuine removal rather than a visual cover, rotate it upright first.
What this does not do, stated as plainly as what it does
Find & Cover removes matched text from a page’s own visible content. It does not currently sweep hidden metadata, comments, attached files, or earlier revisions a document might carry — the roadmap says so directly rather than implying a broader guarantee than the one actually being made. The PDF X-ray tool in this same editor exists specifically to show you what else a file contains, if a full check matters for what you are handling.
Straight answers
Does covering text with a black box count as redaction?
No, and this is the single most common and most damaging mistake in free PDF tools. A shape drawn on top of text changes what the page looks like, not what the page contains — the original text is still there underneath, in the file’s own structure, selectable by anyone who copies and pastes it or opens the file in a different viewer that renders the layers differently. This is not a hypothetical: it has exposed witness names, settlement figures, and classified material in real, documented cases.
How is this different from a highlight or a coloured shape?
Every other cover-like tool in this editor — highlights, shapes, the freehand pen — works exactly the way people assume a black box works: it draws on top, and the original content stays in the file underneath. That is fine for annotation and genuinely wrong for concealment, so none of those tools are labelled as redaction and none of them should be relied on for it. Find & Cover is the one tool in this editor that is actually different: on export, it deletes the matched text from the page’s own content stream. It does not exist anymore in the file, not merely underneath a visual layer.
Can I verify this myself, or do I have to take your word for it?
You do not have to take anyone’s word for it, on this specifically. Redact something, export it, then open the export’s own verification report — it independently re-extracts text from the exported file and states plainly whether the redacted content is actually gone, the same reopen-and-check step every export already goes through for page count and file structure. You can go further yourself: open the exported PDF in any other PDF reader and try to select or search for the text you redacted. It will not be there.
Is there any case where this still does not fully work?
One, and it is stated on the tool itself at the moment it would matter, not buried in a policy page: on a page that has been rotated, the match is still covered visually, but the underlying text is left in place rather than removed. Correctly converting a rotated redaction box back into the page’s own, un-rotated coordinate space is real additional work not yet done — and shipping a version that might remove the wrong region, or remove the right region incorrectly, would be worse than being honest that this one case does not yet get full removal. Rotate the page upright first if a rotated page needs a genuine removal, not just a visual cover.
Does this remove hidden metadata, comments, or earlier revisions too?
Not yet, and this page will say so plainly if that changes rather than let the scope quietly drift. What Find & Cover removes today is the matched text from the page’s own visible content stream. A document can still carry other content worth checking before you consider it fully clean — the PDF X-ray tool in this editor is built specifically to show you what else a file contains: embedded fonts, hidden layers, attached files, and revision history a redaction pass does not touch.